Many attacks against security microcontrollers have the aim of influencing the output of data such that not just permitted data but also other memory contents, such as secret keys or protected program code; are output. In order to achieve what is known as a “dump”, it is possible to manipulate the CPU or a peripheral controller, for example.
In order to counteract attacks, it is possible to provide sensors that can be used to detect the attack conditions. The security microcontroller can then be put into a safe state, for example in which the output of the data is interrupted, in good time. The sensors can for the most part be bypassed by improved attacks, however.
The executing modules or circuit portions of the security microcontroller can also be embodied on a redundant basis in order to detect manipulations on one of the portions during operation. When a manipulation is recognized, the security microcontroller can be put back into a safe state in good time. However, modern and future security microcontrollers have very many components involved in the data transmission. Every single component would need to be stored redundantly in order to safeguard the overall security of the product.